The neural basis of semantics has long been a central topic across disciplines such as linguistics, cognitive neuroscience, and artificial intelligence. Traditional approaches in neuroscience often struggle to isolate semantic processing from syntactic influences without compromising the naturalness of sentence comprehension, due to constraints in experimental paradigms. Meanwhile, large language models (LLMs) have shown remarkable alignment with human brain activity, offering new tools to investigate language representation in the brain through brain encoding techniques. However, extracting purely semantic representations from LLMs remains a significant challenge, as their internal embeddings inherently entangle syntax and semantics. To address this challenge, we introduce a semantic representation isolation method based on large language models. Specifically, we propose Augmented Semantic Extraction (ASE), which conditionally perturbs word order to disrupt syntactic structures, thereby isolating semantic representations from the mixed encodings produced by LLMs. We evaluate ASE using multiple probing techniques and demonstrate that it effectively purifies semantic content from language model representations under a linear decoding framework. Leveraging ASE, we perform brain encoding experiments on a functional magnetic resonance imaging (fMRI) dataset collected during natural spoken narrative comprehension. Our findings reveal the cortical organization of semantic processing. The results provide neural evidence that semantic processing constitutes the dominant component of language comprehension in the brain.

As an advantageous technique and service, the blockchain has shown great development and application prospects. However, its security has also met great challenges, and many security vulnerabilities and attack issues in blockchain-based services have emerged. Recently, security issues of blockchain have attracted extensive attention. However, there is still a lack of blockchain security research from a full-stack architecture perspective, as well as representative quantitative experimental reproduction and analysis. We aim to provide a security architecture to solve security risks in blockchain services from a full-stack architecture perspective. Meanwhile, we propose a formal definition of the full-stack security architecture for blockchain-based services, and we also propose a formal expression of security issues and defense solutions from a full-stack security perspective. We use ConCert to conduct a smart contract formal verification experiment by property-based testing. The security vulnerabilities of blockchain services in the Common Vulnerabilities and Exposures (CVE) and China Nation Vulnerability Database (CNVD) are selected and enumerated. Additionally, three real contract-layer real attack events are reproduced by an experimental approach. Using Alibaba's blockchain services and Identity Mixer in Hyperledger Fabric as a case study, the security problems and defense techniques are analyzed and researched. At last, the future research directions are proposed.

Towards automated retinal screening, this paper makes an endeavor to simultaneously achieve pixel-level retinal lesion segmentation and image-level disease classification. Such a multi-task approach is crucial for accurate and clinically interpretable disease diagnosis. Prior art is insufficient due to three challenges, i.e., lesions lacking objective boundaries, clinical importance of lesions irrelevant to their size, and the lack of one-to-one correspondence between lesion and disease classes. This paper attacks the three challenges in the context of diabetic retinopathy (DR) grading. We propose Lesion-Net, a new variant of fully convolutional networks, with its expansive path redesigned to tackle the first challenge. A dual Dice loss that leverages both semantic segmentation and image classification losses is introduced to resolve the second challenge. Lastly, we build a multi-task network that employs Lesion-Net as a side-attention branch for both DR grading and result interpretation. A set of 12K fundus images is manually segmented by 45 ophthalmologists for 8 DR-related lesions, resulting in 290K manual segments in total. Extensive experiments on this large-scale dataset show that our proposed approach surpasses the prior art for multiple tasks including lesion segmentation, lesion classification and DR grading.

Traditional STEM education models in electrical engineering and computer science rely on structured classes, laboratories, and textbooks to transfer key concepts. Even though this process meets most of the ABET objectives, it does not respond well to current workforce needs that require widely accessible programs that will provide a large pool of graduates with STEM backgrounds, analytical and programming skills, critical thinking, and leadership abilities. In this work in progress paper, we describe our efforts to motivate students to pursue studies in STEM areas. We accomplish this by creating and disseminating modules that demonstrate how math and engineering theory enable modern applications such as those embedded in wireless devices.

The exponential growth of high-throughput DNA sequence data has posed great challenges to genomic data storage, retrieval and transmission. Compression is a critical tool to address these challenges, where many methods have been developed to reduce the storage size of the genomes and sequencing data (reads, quality scores and metadata). However, genomic data are being generated faster than they could be meaningfully analyzed, leaving a large scope for developing novel compression algorithms that could directly facilitate data analysis beyond data transfer and storage. In this article, we categorize and provide a comprehensive review of the existing compression methods specialized for genomic data and present experimental results on compression ratio, memory usage, time for compression and decompression. We further present the remaining challenges and potential directions for future research.

SUMMARY: Exhaustive mapping of next-generation sequencing data to a set of relevant reference sequences becomes an important task in pathogen discovery and metagenomic classification. However, the runtime and memory usage increase as the number of reference sequences and the repeat content among these sequences increase. In many applications, read mapping time dominates the entire application. We developed CompMap, a reference-based compression program, to speed up this process. CompMap enables the generation of a non-redundant representative sequence for the input sequences. We have demonstrated that reads can be mapped to this representative sequence with a much reduced time and memory usage, and the mapping to the original reference sequences can be recovered with high accuracy. AVAILABILITY AND IMPLEMENTATION: CompMap is implemented in C and freely available at http://csse.szu.edu.cn/staff/zhuzx/CompMap/. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: The exponential growth of next generation sequencing (NGS) data has posed big challenges to data storage, management and archive. Data compression is one of the effective solutions, where reference-based compression strategies can typically achieve superior compression ratios compared to the ones not relying on any reference. RESULTS: This paper presents a lossless light-weight reference-based compression algorithm namely LW-FQZip to compress FASTQ data. The three components of any given input, i.e., metadata, short reads and quality score strings, are first parsed into three data streams in which the redundancy information are identified and eliminated independently. Particularly, well-designed incremental and run-length-limited encoding schemes are utilized to compress the metadata and quality score streams, respectively. To handle the short reads, LW-FQZip uses a novel light-weight mapping model to fast map them against external reference sequence(s) and produce concise alignment results for storage. The three processed data streams are then packed together with some general purpose compression algorithms like LZMA. LW-FQZip was evaluated on eight real-world NGS data sets and achieved compression ratios in the range of 0.111-0.201. This is comparable or superior to other state-of-the-art lossless NGS data compression algorithms. CONCLUSIONS: LW-FQZip is a program that enables efficient lossless FASTQ data compression. It contributes to the state of art applications for NGS data storage and transmission. LW-FQZip is freely available online at: http://csse.szu.edu.cn/staff/zhuzx/LWFQZip.

Problem domains are commonly decomposed hierarchically to fully utilize parallel resources in modern microprocessors. Such decompositions can be provided as library routines, written by experienced experts, for general algorithmic patterns. But such APIs tend to be constrained to certain architectures or data sizes. Integrating them with application code is often an unnecessarily daunting task, especially when these routines need to be closely coupled with user code to achieve better performance. This paper contributes HiDP, a high-level hierarchical data parallel language. The purpose of HiDP is to improve the coding productivity of integrating hierarchical data parallelism without significant loss of performance. HiDP is a source-to-source compiler that converts a very concise data parallel language into CUDA C++ source code. Internally, it performs necessary analysis to compose user code with efficient and architecture-aware code snippets. This paper discusses various aspects of HiDP systematically: the language, the compiler and the run-time system with built-in tuning capabilities. They enable HiDP users to express algorithms in less code than low-level SDKs require for native platforms. HiDP also exposes abundant computing resources of modern parallel architectures. Improved coding productivity tends to come with a sacrifice in performance. Yet, experimental results show that the generated code delivers performance very close to handcrafted native GPU code.

This paper develops and evaluates search and optimization techniques for autotuning 3D stencil (nearest neighbor) computations on GPUs. Observations indicate that parameter tuning is necessary for heterogeneous GPUs to achieve optimal performance with respect to a search space. Our proposed framework takes a most concise specification of stencil behavior from the user as a single formula, autogenerates tunable code from it, systematically searches for the best configuration and generates the code with optimal parameter configurations for different GPUs. This autotuning approach guarantees adaptive performance for different generations of GPUs while greatly enhancing programmer productivity. Experimental results show that the delivered floating point performance is very close to previous handcrafted work and outperforms other autotuned stencil codes by a large margin. Furthermore, heterogeneous GPU clusters are shown to exhibit the highest performance for dissimilar tuning parameters leveraging proportional partitioning relative to single-GPU performance.

This paper develops and evaluates search and optimization techniques for auto-tuning 3D stencil (nearest-neighbor) computations on GPUs. Observations indicate that parameter tuning is necessary for heterogeneous GPUs to achieve optimal performance with respect to a search space. Our proposed framework takes a most concise specification of stencil behavior from the user as a single formula, auto-generates tunable code from it, systematically searches for the best configuration and generates the code with optimal parameter configurations for different GPUs. This auto-tuning approach guarantees adaptive performance for different generations of GPUs while greatly enhancing programmer productivity. Experimental results show that the delivered floating point performance is very close to previous handcrafted work and outperforms other auto-tuned stencil codes by a large margin.

Data-parallel languages feature fine-grained parallel primitives that can be supported by compilers targeting modern many-core architectures where data parallelism must be exploited to fully utilize the hardware. Previous research has focused on converting data-parallel languages for SIMD (single instruction multiple data) architectures. However, directly applying them to today's SIMT (single instruction multiple thread) architectures does not guarantee competitive performance. We propose cuNesl, a compiler framework to translate and optimize NESL into parallel CUDA programs for SIMT architectures. By converting recursive calls into while loops, we ensure that the hierarchical execution model in GPUs can be exploited on the "flattened" code. The performance gap between our auto-generated CUDA code and hand-crafted CUDA code thus narrows while programmability is greatly increased. Our compiler outperforms handwritten parallel code running on CPUs in terms of both execution time and programmability.

Emerging accelerating architectures, such as GPUs, have proved successful in providing significant performance gains to various application domains. However, their viability to operate on general streaming data is still ambiguous. In this paper, we propose GStream, a general-purpose, scalable data streaming framework on GPUs. The contributions of GStream are as follows: (1) We provide powerful, yet concise language abstractions suitable to describe conventional algorithms as streaming problems. (2)We project these abstractions onto GPUs to fully exploit their inherent massive data parallelism.(3) We demonstrate the viability of streaming on accelerators. Experiments show that the proposed framework provides flexibility, programmability and performance gains for various benchmarks from a collection of domains, including but not limited to data streaming, data parallel problems and numerical codes.

Document clustering plays an important role in data mining systems. Recently, a flocking-based document clustering algorithm has been proposed to solve the problem through simulation resembling the flocking behavior of birds in nature. This method is superior to other clustering algorithms, including k-means, in the sense that the outcome is not sensitive to the initial state. One limitation of this approach is that the algorithmic complexity is inherently quadratic in the number of documents. As a result, execution time becomes a bottleneck with large number of documents. In this paper, we assess the benefits of exploiting the computational power of Beowulf-like clusters equipped with contemporary Graphics Processing Units (GPUs) as a means to significantly reduce the runtime of flocking-based document clustering. Our framework scales up to over one million documents processed simultaneously in a sixteen-node moderate GPU cluster. Results are also compared to a four-node cluster with higher-end GPUs. On these clusters, we observe 30X-50X speedups, which demonstrate the potential of GPU clusters to efficiently solve massive data mining problems. Such speedups combined with the scalability potential and accelerator-based parallelization are unique in the domain of document-based data mining, to the best of our knowledge.

In this paper, we improve the soft-threshold by using Markov random field (MRF) theory to preserve the edges of images. According to MRF theory, the similarity problem of the edge structure can be transformed into an energy function minimization problem. Through computing all the energies of the edge structures in clique c, we can find the edge structure with the minimum energy, and regard it as the edge structure of this place. In this way, we can gain the edge information of an image. Here we detect the edge of image in wavelet domain. In our method we preserve the wavelet coefficients which belong to edge set; for the other wavelet coefficients, the soft threshold method is applied. Then, we reconstruct the treated wavelet coefficients by using inverse wavelet transform. The experiment result shows that our method is better than the hard threshold and soft threshold method. It will not only smooth away noise, but also preserve the edge of image.

A blind MMSE multipath combining method for a RAKE receiver is proposed. The method make use of the data before and after the onset of the desired user's transmission. By performing generalized eigendecomposition of the autocorrelation matrices of the post-RAKE signals in the two interval, an MMSE combiner for the desired user signal can be obtained without the need of the channel state information. The method does not require a precise model nor assume whiteness of the interference. An adaptive algorithm is also developed.

A blind MMSE multipath combining method for RAKE receivers in long-code CDMA systems is proposed. The method uses the data before and after the onset of the desired user's transmission. It shows that the generalized eigendecomposition of the autocorrelation matrices of the post-RAKE signals in these two intervals gives an MMSE combiner for the desired user signal, without the need of the channel state information. The method does not require precise models of the interference but still offers similar performance as that of the optimal RAKE combiner or even better under certain circumstances. An adaptive algorithm is also developed.